Tootfinder

Learning to Diagnose Privately: DP-Powered LLMs for Radiology Report Classification
Payel Bhattacharjee, Fengwei Tian, Ravi Tandon, Joseph Lo, Heidi Hanson, Geoffrey Rubin, Nirav Merchant, John Gounley
https://arxiv.org/abs/2506.04450

Learning to Diagnose Privately: DP-Powered LLMs for Radiology Report Classification
Purpose: This study proposes a framework for fine-tuning large language models (LLMs) with differential privacy (DP) to perform multi-abnormality classification on radiology report text. By injecting calibrated noise during fine-tuning, the framework seeks to mitigate the privacy risks associated with sensitive patient data and protect against data leakage while maintaining classification performance. Materials and Methods: We used 50,232 radiology reports from the publicly available MIMIC-CXR …

Verdad que hay fútbol de selección hoy. Una buena oportunidad para ver lucirse a la Argentina campeona del mundo.
Mi tradicional pronóstico:
Chile 0 - 4 Argentina
#ChilexArgentina #clasificatoriasconmebol

Domain Adaptation Method and Modality Gap Impact in Audio-Text Models for Prototypical Sound Classification
Emiliano Acevedo, Mart\'in Rocamora, Magdalena Fuentes
https://arxiv.org/abs/2506.04376

Domain Adaptation Method and Modality Gap Impact in Audio-Text Models for Prototypical Sound Classification
Audio-text models are widely used in zero-shot environmental sound classification as they alleviate the need for annotated data. However, we show that their performance severely drops in the presence of background sound sources. Our analysis reveals that this degradation is primarily driven by SNR levels of background soundscapes, and independent of background type. To address this, we propose a novel method that quantifies and integrates the contribution of background sources into the classifi…

Schafft KI die Entwickler*innen ab? Oder ist das Clickbait?
Meine Juli-Kolumne im @… , in der ich finde, dass es eigentlich eine gute Sache ist, wenn uns Maschinen Arbeit abnehmen.
Um die neue Buzzwords Deregulierung und Bürokratieabbau geht's auch und ihr ahnt schon, warum.
Künstliche Intelligenz – Weniger Arbeit durch KI – eigentlich eine …

#photography #bloomScrolling #iris

This https://arxiv.org/abs/2405.12717 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_mat…

A note on the Thom morphism for the classifying space of certain Lie groups and gauge groups
We give a complete description of which non-torsion generators are not in the image of the Thom morphism from complex cobordism to integral cohomology for the classifying space of exceptional Lie groups except for E_8. We then show that the Thom morphism is not surjective for the classifying space of the gauge group of a principal E_7-bundle over the four-dimensional sphere. We use the results to detect nontrivial elements in the kernel of the reduced Thom morphism for Lie groups and their clas…

Classification and enumeration of solid-solid phase transition mechanisms
Fang-Cheng Wang, Qi-Jun Ye, Yu-Cheng Zhu, Xin-Zheng Li
https://arxiv.org/abs/2506.05105

Classification and enumeration of solid-solid phase transition mechanisms
Crystal-structure match (CSM), the atom-to-atom correspondence between two crystalline phases, is used extensively to describe solid-solid phase transition (SSPT) mechanisms. However, existing computational methods cannot account for all possible CSMs. Here, we propose a formalism to classify all CSMs into a tree structure, which is independent of the choices of unit cell and supercell. We rigorously proved that only a finite number of noncongruent CSMs are of practical interest. By representin…

webkb: WebKB graphs (1998)
Web graphs crawled from four Computer Science departments in 1998, with each page manually classified into one of 7 categories: course, department, faculty, project, staff, student, or other. All graphs included in a single .zip; also included are 'co-citation' graphs, which links i and j if they both point to some k. Edge weights count the number of links from i to j.
This network has 349 nodes and 1393 edges.
Tags: Informational, Web gra…

via @…

EMBER2024 -- A Benchmark Dataset for Holistic Evaluation of Malware Classifiers
Robert J. Joyce, Gideon Miller, Phil Roth, Richard Zak, Elliott Zaresky-Williams, Hyrum Anderson, Edward Raff, James Holt
https://arxiv.org/abs/2506.05074

EMBER2024 -- A Benchmark Dataset for Holistic Evaluation of Malware Classifiers
A lack of accessible data has historically restricted malware analysis research, and practitioners have relied heavily on datasets provided by industry sources to advance. Existing public datasets are limited by narrow scope - most include files targeting a single platform, have labels supporting just one type of malware classification task, and make no effort to capture the evasive files that make malware detection difficult in practice. We present EMBER2024, a new dataset that enables holisti…